Feeding Characteristics of a Japanese Pitviper, Title okinavensis, on Okinawa Island: Seasonally Biased but Ontogenetically Stable Exploitation on Small Frogs

Author(s) MORI, Akira; TODA, Mamoru

Citation Current Herpetology (2011), 30(1): 41-52

Issue Date 2011-07

URL http://hdl.handle.net/2433/197271

Right © 2011 The Herpetological Society of

Type Journal Article

Textversion publisher

Kyoto University Current Herpetology 30(1): 41–52, June 2011 © 2011 by The Herpetological Society of Japan

Feeding Characteristics of a Japanese Pitviper, , on Okinawa Island: Seasonally Biased but Ontogenetically Stable Exploitation on Small Frogs

1 1,2 AKIRA MORI * AND MAMORU TODA

1 Department of Zoology, Graduate School of Science, Kyoto University, Sakyo, Kyoto 606–8502, JAPAN 2 Present address: Tropical Biosphere Research Center, University of the Ryukyus, Nishihara, Okinawa 903–0213, JAPAN

Abstract: Ovophis okinavensis is an endemic pitviper of the subtropical region of Japan. Based on information from various localities, the has been considered a dietary generalist that exploits various terrestrial . We analyzed stomach contents of O. okinavensis in a northern mountain area on Okinawa Island, based on data of a 10-years-study, to examine its trophic features at the population level. We recovered a total of 461 prey items, which were comprised of six, three, five, and three species of frogs, , birds, and mammals, respectively. Approximately 93% of dietary items consisted of frogs, the majority of which belonged to two explosively breeding species aggregating to restricted spots along the stream only during their reproductive season (winter). On the other hand, from April to November, the mainly depended on vertebrates other than frogs. Neither an ontogenetic dietary shift nor intersexual dietary divergence was evident although males tended to exploit one of the two explosively breeding frog species more frequently than females. There were weak positive correlations between snake size and prey size, but large did not drop small prey items from their diet. Average body mass of each prey item was only 5.2% of snake body mass. Characteristics of trophic features of O. okinavensis of this population, such as extremely high dependency on frogs in winter, the absence of ontogenetic dietary shift, and utilization of small sized prey, are discussed in comparison with other viperids.

Key words: Crotalinae; Prey size; Ontogenetic dietary shift; Ovophis okinavensis; Sexual difference

INTRODUCTION exclusive of Australasia and Antarctica (Pough et al., 2004). Since those early works by a few is one of the much diversified authors, such as Saint Girons (1952), Fitch snake families widely radiating over the world (1960), and Klauber (1972), the trophic ecol- ogy and feeding characteristics of viperid * Corresponding author. Tel: +81–75–753–4075; snakes have been relatively well studied chiefly Fax: +81–75–753–4075; on the basis of data from European and North E-mail address: [email protected] American species. With respect to members 42 Current Herpetol. 30(1) 2011 of the family from other regions, most compa- okinavensis, such as sexual differences, rable studies started much more recently ontogenetic shift, seasonal change, and effects (South America: Valdujo et al., 2002; Nogueira of snake’s body size on the prey properties. et al., 2003; Monteiro et al., 2006, Central Here, we investigate the feeding habit of O. Africa: Luiselli and Akani, 2003; Ineich et al., okinavensis by analyzing sexual, ontogenetic, 2006, Southeast : Daltry et al., 1998; and seasonal divergences in prey taxa and size, Creer et al., 2002; Mori et al., 2002; Shine and based on a long-term study conducted in the Sun, 2003; Lin and Tu, 2008), but these studies northern mountain area on Okinawa Island. also have revealed their diverse feeding habits We then discuss the trophic features of this as ranging from frog specialists (Ineich et al., pitviper by comparing its general feeding 2006) to generalists (Creer et al., habits with those of other viperid snakes 2002; Valdujo et al., 2002; Monteiro et al., previously reported. 2006). These studies, as well as several other studies on some North American species, have MATERIALS AND METHODS also highlighted a possible common tendency of ontogenetic dietary shift, typically from The diet of O. okinavensis was examined on ectothermic to endothermic prey (Klauber, the basis of the stomach contents of individual 1972; Campbell and Lamar, 1989; Daltry et snakes captured from December 1996 to al., 1998; Shine and Wall, 2007; Lin and Tu, February 2006 in a limited area of the north- 2008), along with improvement of ability to ern part of Okinawa Island, Ryukyu Archipel- ingest larger prey through structural modifica- ago, Japan. The main study area (ca. 25 ha) tions of trophic apparatus, an evolutionary encompasses the upper streams of Zatsun innovation characteristic to viperid snakes River and surrounding hilly environments of (Pough and Groves, 1983; Greene, 1992; but Yambaru. Vegetation was dominated by see Cundall and Deufel, 2006). Nonetheless, a primary and well-recovered secondary forests great majority of the Asian and African consisting of humid-subtropical broad-leaved viperids remains to be studied (Ineich et al., evergreen trees such as Castanopsis sieboldii. 2006). Furthermore, dietary information on a Meteorological data of the study region is given Asian or African species, even when shown in Mori et al. (2002). Field study was available, mostly derives from museum speci- made throughout the year but with particular mens originated from various localities: very intensities from December to March for the few population-based studies have been con- purpose of investigating the relationship of a ducted despite their obvious importance for winter activity of the snake with the breeding our better understanding of feeding traits in an activity of two species of frogs, Rana sp. evolutionary framework (Thompson, 2005). (formerly erroneously referred to as R. Ovophis okinavensis is a small, subtropical okinavana, see Matsui [2007]; hereafter pitviper distributed on islands in the central referred to as Rana sp. A) and R. narina, both part of the Ryukyu Archipelago, Japan. Cur- breeding in the stream (Mori et al., 2002, rently available dietary records of wild O. 2009). Basically, we haphazardly walked at okinavensis cover all terrestrial vertebrate night along streams and trails, which run classes (Nakachi, 1991; Mori et al., 2002; Toda though the study area irregularly, searching for et al., 2003; Kadota, 2006, 2011; see Mori and snakes on the surface. Whenever we found Moriguchi, 1988 for other references), imply- individual snakes, we collected them except for ing that this species is a generalist feeder. On the cases specified below. We also searched the other hand, despite the abundance of between rocks, insides of crevices of ledges, dietary records, few studies have examined and in burrows on banks. patterns and causes of variations in feeding In addition to the above surveys, we also characteristics within a single population of O. used a road cruising technique to find snakes MORI & TODA—FEEDING CHARACTERISTICS OF PITVIPER 43 on the road to collect dietary data from differ- scale clipping and by painting numbers on ent . The road is an approximately dorsal surfaces of head and posterior body for 7 km segment of the routes Benoki, Ie, and permanent and temporal visual identifications, Chinufuku, which partially surrounds the respectively. In addition, from August 1998, main study area and crosses the Yambaru small (11.5×2.2 mm) passive integrated tran- forests. These routes run through various sponder (PIT) tags were injected under the habitats, such as forests along hill ridges, skin for individual identification. Each snake slopes, and streams, and a forest park with was released at the site of capture basically more open habitats. The straight distance within 24 hours after collection. from the main study area to the farthest point When we found snakes that had been of the segment is approximately 3 km. The collected and released within a few weeks width of the paved road varied from 5 to 8 m. before (visually determined by checking the We drove at 20 to 40 km/h, depending on painting), we usually did not collect them or visibility, and collected snakes whenever we did not examine stomach contents so as not to found them on the way to and back from the disturb them too frequently. Prey main study area. confirmed by direct observations on predation Snakes were brought to the nearby field were also recorded as food items. station within 12 hours of collection, where In our earlier study, only a few snakes were they were measured for snout-vent length found along the stream by visual census (SVL), body mass (BM), maximum head width between late spring and late fall (Mori et al., (MHW), interocular distance (IOD: distance 2002). We therefore periodically examined between exterior edges of eyes), and upper jaw between March 2004 and February 2005 length (JL: straight distance between the tip of stomach contents of four radio-tracked snakes the snout and the posterior edge of the eighth (two males and two females), to which we supralabial). MHW and JL were measured surgically implanted radio tags (Holohil SB- while the snakes were gently held by the neck 2T, 5 g) intraperitoneally following the proce- region so as to keep the head slightly above dures described by Reinert and Cundall (1982) the substrate and the mouth closed. Although and Nishimura et al. (1995). These snakes JL is not a direct measurement of the upper were initially collected along the stream and jaw, we used this measurement as one of the released at the site of capture four to six days approximate indices of head size. Sex was after surgery. We located the radio-tracked determined by using a probe, by everting hemi- snakes once every two weeks and collected penes, or by examining the external shape of them to examine the presence of the stomach the tail base. Stomach contents were exam- contents unless they stayed away from our ined by palpation and forced-regurgitation. access, such as under large rocks or in deep Recovered prey items were identified and crevices. If the snakes had stomach contents, measured for MHW, SVL, and BM whenever we processed them in the same way as described possible. Direction of ingestion of each prey above. The original data of stomach contents item was determined either as head first, hind collected between December 1996 and March (tail or hindleg[s]) first, or midbody first (swal- 2000 are the same as those presented in Mori lowing by bending at the midbody of the prey). et al. (2002). When the prey items were frogs, they were Sexual differences of SVL and total prey basically fed back to the stomach after identifi- body mass were examined using Welch t-test cation and measurements. Unidentified items because of the heterogeneity of variance. and animals other than frogs were preserved Relative head size was compared between the in 10% formalin for later identification and sexes by ANCOVA using log transformed data measurements. The snakes captured and with SVL as the covariate. Student t-test and examined were individually marked by ventral ANOVA were used for other comparisons of 44 Current Herpetol. 30(1) 2011 means between two groups and among more than two groups, respectively. Differences in diet composition between the sexes and among seasons were examined using Fisher’s exact probability test. Correlations of snake size and prey size were examined using Kendall rank correlation. For the analyses of snake body measurements, only the data for each individual at first capture were used so as to ensure the statistical independence of each data set. All statistical analyses were conducted using StatView (ver. 5.0, SAS, Inc., Cary, NC; http://www.statview.com) with a significant level of P=0.05.

RESUTLS

Females were significantly larger in SVL than males (Welch t-test, df=1, 257.16, t=7.21, P<0.0001: female, x±SE=481±6.34 mm; male, 430±3.5 mm). ANCOVA revealed signifi- cant sexual differences in slope of regression lines between SVL and MHW (df=1, 366, F= 23.71, P<0.0001). Females have relatively wider head than males, and this difference FIG. 1. Comparisons of maximum head width becomes clearer as the body size of snakes against snout-vent length (A) and jaw length (B) becomes larger (Fig. 1A). There were no between male (open circles) and female (closed significant sexual differences in either slope or circles) Ovophis okinavensis. elevation in regression lines between SVL and IOD (slope, df=1, 366, F=0.41, P=0.847; vs. other three taxa combined, Fisher test, elevation, df=1, 366, F=0.25, P=0.620) and P=0.173). However, when only those two between SVL and JL (slope, df=1, 366, F= predominant frog species were considered, 0.82, P=0.366; elevation, df=1, 366, F=0.41, males had Rana sp. A more frequently than P=0.525). There were significant sexual females, and the latter had R. narina more differences in slope in regression lines between frequently than the former (P<0.0001). JL and MHW (df=1, 366, F=40.24, P< In January-March approximately 30% of 0.0001). Females showed relatively wider and snakes had stomach contents (Table 2). In shorter head than males (Fig. 1B). contrast, only 4.3% and 11.1% of females and A total of 461 prey items, consisting of six males, respectively, had stomach contents in species of frogs, one species of lizards, two the period from April to November. In species of snakes, five species of passerine December nearly half of males had stomach birds, and one species each of , shrews, contents, whereas no more than 14.4% of and bats, were identified (Table 1). Nearly females had prey in stomachs. Both males and 93% of the prey items were frogs, and the females predominantly depended on frogs in majority (85.9%) of them consisted of two January-March and December (Fig. 2). Between species, R. narina and Rana sp. A. There April and November, the snakes more fre- were no significant differences in the frequency quently depended on other vertebrate taxa of frogs in food items between the sexes (frogs (reptiles, birds, and mammals) than in the MORI & TODA—FEEDING CHARACTERISTICS OF PITVIPER 45

TABLE 1. Stomach contents of Ovophis okinavensis. Numerals in table indicate the individual numbers of prey items followed by ratios (%) in parentheses.

Taxon Male Female Total Amphibia Anura Rana narina 59 (18.6) 93 (64.6) 152 (33.0) Rana sp. A* 220 (69.4) 24 (16.7) 244 (52.9) Rana ishikawae 1 (0.3) 2** (1.4) 3 (0.7) Rana holsti 1 (0.3) 0 1 (0.2) Rhacophorus viridis 5 (1.6) 1 (0.7) 6 (1.3) Buergeria japonica 1 (0.3) 0 1 (0.2) Unidentified frog 11 (3.5) 10 (7.0) 21 (4.6) Reptilia Lacertilia Japalura polygonata 1 (0.3) 1 (0.7) 2 (0.4) Serpentes Achalinus werneri 1 (0.3) 0 1 (0.2) Amphiesma pryeri 0 1 (0.7) 1 (0.2) Aves Passeriformes Anthus hodgsoni 0 1 (0.7) 1 (0.2) Cettia diphone 2 (0.6) 0 2 (0.4) Erithacus cyanurus 1 (0.3) 1 (0.7) 2 (0.4) Luscinia calliope 1 (0.3) 0 1 (0.2) Turdus pallidus 1 (0.3) 0 1 (0.2) Unidentified bird 0 2 (1.4) 2 (0.4) Mammal Insectivora Crocidura watasei 10 (3.2) 7 (4.9) 17 (3.7) Rodentia Rattus rattus 0 1 (0.7) 1 (0.2) Rattus sp. 1 (0.3) 0 1 (0.2) Chiroptera Murina ryukyuana 1 (0.3) 0 1 (0.2) Total 317 144 461 * Formerly referred to as Rana okinavana (see Matsui, 2007). ** One of these frogs was found just grasped by a snake, but eventually the snake released the frog.

TABLE 2. Seasonal changes in the frequency (%) 87.3%, female, 74.1%), this difference was of snakes with stomach contents. Numerals in each set not statistically significant (Fisher test, P= of parentheses are the number of snakes with stomach 0.077). contents/the number of snakes examined. In both males and females, there were no Season clear alterations in diet with growth (Fig. 3). Sex All four classes of terrestrial vertebrates Jan-Mar Apr-Nov Dec (amphibians, reptiles, birds, and mammals) Female 31.3 4.3 14.4 seemed to be fairly evenly distributed in the (90/288) (2/46) (15/104) diet over the size ranges of snakes. Nonethe- Male 26.6 11.1 48.4 less, mean SVL of snakes that had frogs was (66/248) (6/54) (133/275) slightly but significantly larger than that of Total 29.1 8.0 39.1 snakes with mammals in their stomach (ANOVA, df=3, 476, F=3.918, P<0.01; Tukey-Kramer (156/536) (8/100) (148/379) test, P<0.05; frogs, x±SE=474.8±2.4,4 mam- mals, 441.0±12.6; males and females were other two seasons (Fisher test, both P< combined because of small sample size of 0.0001). Although males tended to have more reptiles and birds). In both sexes, no significant Rana sp. A than females in December (male, differences were recognized in SVL between 46 Current Herpetol. 30(1) 2011

snakes that had R. narina and Rana sp. A in their stomach (t-test, male, df=119, t=0.491, P=0.624; female, df=61, t=1.314, P=0.194). There was no significant correlation between snake SVL and prey SVL (Kendall τ=0.098, P=0.077: Fig. 4A). Snakes smaller than 380 mm in SVL tended to eat only small prey (around 40 mm in SVL), but larger snakes sometimes consumed even smaller prey. The ranges of SVL of prey items consumed by snakes larger than 400 mm seemed to be nearly constant, and there were no clear sexual differences in the size range of prey items. On the other hand, there was a significant positive correlation between snake MHW and prey head width (τ=0.346, P<0.0001; Fig. 4B). Nonetheless, there were no tendencies to drop IG F . 2. Seasonal changes in the composition of small-sized prey from the diet of large snakes. prey items of male and female Ovophis okinavensis. There was a significant, but weak positive Numerals above columns are sample sizes. correlation between snake BM and prey BM (τ=0.193, P=0.0012; Fig. 5A). Prey mass consumed by each snake ranged from 1.5 to 28.1% of the snake’s body mass with mean of 5.2% (SD=4.4). There were no significant differences in relative prey width (prey head width/snake MHW) between prey swallowed head first and hind first (df=145, t=1.51, P=0.134; head first, x±SE=0.529±0.011,4 hind first, 0.555±0.012; Fig. 4B). Multiple prey items were often found in a single snake stomach. Maximum number of prey items per stomach was nine, and approxi- mately 44% (107/244) of snakes had more than one prey items. Body mass of partially digested frogs (R. narina and Rana sp. A) with an intact tibia bone was estimated using regression lines between tibia length and body mass obtained from undigested frogs (males and females of each species were treated sepa- rately), and total prey body mass (TPM) was calculated for each snake. TPM varied from 2.4 to 66.5 g and was significantly, but weakly correlated with snake BM (Kendall τ=0.201, FIG. 3. Relationships between snout-vent length P=0.0007; Fig. 5B). There were significant dif- and prey taxa in female (A) and male Ovophis ferences in TPM between males and females okinavensis (B). (Welch t-test, df=1, 47.65, t=2.32, P=0.0249; males, 12.8±1.0 g, female, 19.7±2.8 g). TPM consumed by each snake varied from 1.5 to MORI & TODA—FEEDING CHARACTERISTICS OF PITVIPER 47

FIG. 4. Relationships between size of Ovophis FIG. 5. Relationships of body mass of Ovophis okinavensis and that of prey items. A, snout-vent okinavensis with that of each prey item (A) and length of male (open circles) and female (closed total body mass of prey items from each stomach circles) snakes vs. that of prey. B, head width of (B). Lines in A and B show positions where individ- snake vs. head width of prey animals that were ual prey mass is equal to 10% of snake body mass swallowed hind first (open circles), head first (closed and where the total prey body mass is equal to 20% triangles), and midbody first (crosses). The line of snake body mass, respectively. Open circles, male; shows the position where prey head width is equal closed circles, female. to snake head width. to aerial bats and birds. This evidence shows 50.5% of snake BM. There were no signifi- that O. okinavensis is, at least qualitatively, a cant differences in relative TPM (TPM/snake generalist feeder. However, vast majority of BM) between the sexes (t-test, df=127, t= food items (86%) were comprised of only two 0.947, P=0.345; male, 12.6±0.8%; female, species of frogs, R. narina and R. sp A, indi- 11.0±1.7%). cating that snakes in this population heavily depend on frogs. This high dependency on DISCUSSION frogs, particularly between December and March, is obviously due to the exploitation of Ovophis okinavensis in the present study aggregated frogs that explosively breed at area exploited 17 species of animals that cover restricted sites along streams in winter (also all four extant terrestrial vertebrate classes see Mori et al., 2002, 2009). Possible advan- occupying diverse ecological niches, from semi- tages of temporarily and spatially limited, but fossorial snakes, through semiaquatic frogs, abundant food resource, such as higher growth terrestrial snakes and rodents, arboreal lizards, rate, better body condition, and longer life 48 Current Herpetol. 30(1) 2011 span, would be worthy for pursuing in a future tative studies are needed to discuss possible study. feeding adaptation of Asian pitvipers that A presumed key innovation in the evolution depend predominantly on frogs. of viperid snakes is the improved ability of Another common trophic feature of vipers is ingesting large prey, which was accomplished the ontogenetic shift from ectothermic to by structural modifications of jaws and related endothermic prey (Mushinsky, 1987; Greene, architectures (Pough and Groves, 1983; Greene, 1997; Shine and Wall, 2007). For example, 1992; but see Cundall and Deufel, 2006). In most members of Crotalus reportedly shift this respect, O. okinavensis seems to provide diet from lizards to mammals (Klauber, 1972; an exceptional case. Although individual O. Campbell and Lamar, 1989). Some Asian okinavensis often had multiple prey items (up pitvipers also show similar ontogenetic dietary to nine frogs) simultaneously, mean BM of shifts (Protobothrops flavoviridis, Nishimura each prey item was 5.2% of snake BM, which et al., 1991; Calloselasma rhodostoma, Dal- is much smaller than that of other vipers and try et al., 1998; gracilis, Lin and even snakes of other families (Greene, 1992; Tu, 2008). Many species of Bothrops, which Sazima, 1992; Martins et al., 2002; Rodríguez- are dietary generalists, drop ectothermic ani- Robles, 2002; Glaudas et al., 2008). However, mals from their diet with growth (Martins et this may not necessarily indicate inability of O. al., 2002). With respect to O. okinavensis, our okinavensis in swallowing large prey. A recent result may reflect the tendency opposite to the report on predation by O. okinavensis observed general viperid pattern, because snakes that in the field near our study site showed that a had frogs in the stomach were slightly larger juvenile O. okinavensis swallowed a frog, whose in SVL (ca. 34 mm) than those that had BM was 49.4% of its predator’s BM (Masunaga mammals. Even so, however, body size of the et al., 2008). Thus, the consumption of rela- snakes largely overlapped among prey catego- tively small prey by O. okinavensis in the ries, leaving the tendency of ontogenetic present study site may simply reflect limitation dietary shift rather obscure. In the analysis of of the size of locally available prey animals. It generalist species of the Latin American is interesting to note that, although anuroph- Bothrops, Martins et al. (2002) indicated that agy in snakes is often associated with the evo- the species that do not show ontogenetic lution of enlargement of feeding apparatus to dietary shifts are small-sized (mean adult total increase gape size to enable swallowing of length=593–742 mm). The absolute small body bulky frogs (Vincent et al., 2006; Mori and size of O. okinavensis that would decrease the Vincent, 2008), O. okinavensis, which pre- possible size range of prey may be partially dominantly exploits frogs, does not seem to responsible for the lack of clear ontogenetic face such a functional challenge (see also Vin- dietary shift (Shine and Wall, 2007). In addi- cent and Mori, 2008) and does not enjoy the tion, the high dependency on locally abundant advantage of the presumed viperid character- frogs throughout almost whole life history istic for swallowing large prey. High depen- may contribute to obscure the possible dietary dency on frogs has also been reported in some shifts, that is, specialization on relatively small populations of Viridovipera stejnegeri in prey may prevent an ontogenetic shift of (Creer et al., 2002) and several species feeding (see Shine and Wall, 2007). Studies on of Trimeresurus (sensu lato) in eastern and O. okinavensis from other localities with southeastern Asia (Orlov et al., 2002), but different prey availability are desired to exam- their size relationships are yet unknown. On ine whether the lack of clear ontogenetic the other hand, species of Causus, a basal dietary shift mentioned above is a specific group of viperids, have been characterized by characteristic of O. okinavensis or it reflects a anurophagy and the consumption of massive more limited local phenomenon. prey items (Ineich et al., 2006). More quanti- Although weak positive correlations were MORI & TODA—FEEDING CHARACTERISTICS OF PITVIPER 49 found, as in some comparable studies on other frogs (Table 1), and there seems to be no snakes (Shine, 1991a; Arnold, 1993), between relationship between the size of the snake and the snake and prey sizes (represented by the the tendency of eating these frogs (Fig. 3). maximum width or mass), there were no clear Thus, it is unlikely that this sexual dietary tendencies to increase the lower prey-size limit divergence is caused by natural selection acting as snake size becomes larger. This results in an to reduce intersexual competition for food ontogenetic telescope pattern (after Arnold, resources. In fact, a proximate factor that 1993) in a relationship between snake size and caused this intersexual difference is obviously a prey size. Although this pattern has been highly biased appearance of males in one considered less common in snakes in general specifically restricted breeding site of Rana sp. (Arnold, 1993; Glaudas et al., 2008), it has A, where few females were observed during the been known in several vipers including four short breeding period of this frog (Mori and generalist species of Bothrops (Martins et al., Toda, unpublished data). Detailed investigation 2002). Continuous exploitation of small prey of intersexual differences of foraging behavior by larger O. okinavensis may be due to the and other possibly relevant ecological factors high dependency on the abundant frog (Rana are needed to elucidate why females of O sp. A). Arnold (1993) indicated that, from the okinavensis neglect such a highly rich food viewpoint of optimal foraging theory, prey resource. selection is determined by the probabilities of encounter, capture, and ingest a prey ACKNOWLEDGEMENTS and costs of these phases. In the case of O. okinavensis in the present study area, the high We are grateful to T. Arioka, M. Azuma, T. availability of the small-sized prey (i.e., Rana Haramura, M. Hasegawa, T. Hayashi, Y. sp. A) and relative easiness of capturing and Hongo, I. Ikeuchi, K. Inamori, R. Ito, S. handling the frog, as well as the ability to Iwanaga, Y. Kadota, H. Kaneda, N. Kidera, T. efficiently swallow small animals in vipers Kitamura, T. Maenosono, G. Masunaga, T. (Pough and Groves, 1983) and the low energy Mizuta, Y. Mori, D. Muramatsu, A. Murata, costs of striking, handling, and ingesting prey E. Nagata, H. Nagatomo, A. Nakachi, M. in snakes in general (Feder and Arnold, 1982; Okuyama, T. Okamoto, H. Randriamahazo, Cruz-Neto et al., 1999, 2001), may be reasons S. Sakata, R. Sasaki, H. Sato, H. Takahashi, why the snake retains small animals as valu- M. Takeshima, I. Takiguchi, K. Tanaka, M. able prey even after growing large. Toda, R. Yamamoto, and T. Yamazaki for Sexual dimorphism in head size has been their assistance in the field. We also thank M. demonstrated in many groups of snakes Motokawa and M. Kajita for their help in including vipers (Shine, 1991b; Vincent and identifying stomach contents. Special thanks Herrel, 2007) and is usually considered to be a are due to H. Ota for his collaboration during result of natural selection that leads to inter- the early stage of this long-term study. This sexual dietary divergence (Shine, 1991b). The study was partially supported by a Grant-in- male biased feeding on Rana sp. A and the Aid for Encouragement of Young Scientists female biased feeding on R. narina may seem- from the Japan Ministry of Education, Science, ingly coincide with this general explanation Sports, and Culture (10740358), a Research because female O. okinavensis, which has Opportunity Grant (1996) from Tropical Bio- larger SVL and larger MHW relative to SVL sphere Research Center, University of the than males, more frequently exploits the larger Ryukyus, a Grant from the Japan Society for species of frog (R. narina). However, the the Promotion of Science (Scientific Research other head size measurements, JL and IOD, C: 18570019), and Grants for the 21st Century did not differ between the sexes. Actually, COE Program (A14) and the Global COE both males and females did eat both species of Program (A06) to Kyoto University. 50 Current Herpetol. 30(1) 2011

LITERATURE CITED ioral context for pitviper evolution. p. 107–117. In: J. A. Campbell and E. D. Brodie, Jr. (eds.), ARNOLD, S. J. 1993. Foraging theory and prey-size— Biology of the Pitvipers. Selva, Tyler. predator-size relations in snakes. p. 87–115. In: GREENE, H. W. 1997. Snakes. The Evolution of R. A. Seigel and J. T. Collins (eds.), Snakes. Mystery in Nature. University of California Press, Ecology & Behavior. McGraw-Hill Inc., New Berkeley. York. INEICH, I., BONNET, X., SHINE, R., SHINE, T., CAMPBELL, J. A. AND LAMAR, W. W. 1989. The BRISCHOUX, F., LEBRETON, M., AND CHIRIO, L. Venomous Reptiles of Latin America. Cornell 2006. What, if anything, is a ‘typical’ viper? University Press, Ithaca. Biological attributes of basal viperid snakes (genus CREER, S., CHOU, W.-H., MALHOTRA, A., AND Causus Wagler, 1830). Biological Journal of the THORPE, R. S. 2002. Offshore insular variation Linnean Society 89: 575–588. in the diet of the Taiwanese bamboo viper KADOTA, Y. 2006. An observation on the stomach Trimeresurus stejnegeri (Schmidt). Zoological contents of the hime-habu, Trimeresurus oki- Science 19: 907–913. navensis, from the northern part of Okinawa CRUZ-NETO, A. P., ANDRADE, D. V., AND ABE, A. Island, Ryukyu Archipelago. Akamata No. 17: S. 1999. Energetic cost of predation: Aerobic 3–4. (in Japanese) metabolism during prey ingestion by juvenile KADOTA, Y. 2011. Is Ovophis okinavensis active rattlesnakes, Crotalus durissus. Journal of only in the cool season? Temporal foraging Herpetology 33: 229–234. pattern of a subtropical in Okinawa, CRUZ-NETO, A. P., ANDRADE, D. V., AND ABE, A. Japan. Zoological Studies 50: 269–275. S. 2001. Energetic and physiological correlates KLAUBER, L. M. 1972. Rattlesnakes. Second Edition. of prey handling and ingestion in lizards and snakes. University of California Press, Berkeley. Comparative Biochemistry and Physiology A. 128: LIN, C.-F. AND TU, M.-C. 2008. Food habits of 515–533. the Taiwanese mountain pitviper, Trimeresurus CUNDALL, D. AND DEUFEL, A. 2006. Influence of gracilis. Zoological Studies 47: 697–703. the delivery system on intraoral prey LUISELLI, L. AND AKANI, G. C. 2003. Diet of transport in snakes. Zoologischer Anzeiger 245: sympatric gaboon vipers (Bitis gabonica) and 193–210. nose-horned vipers (Bitis nasicornis) in southern DALTRY, J. C., WÜSTER, W., AND THORPE, R. S. Nigeria. African Journal of Herpetology 52: 101– 1998. Intraspecific variation in the feeding ecol- 106. ogy of the crotaline snake Calloselasma rhodos- MARTINS, M., MARQUES, O. A. V., AND SAZIMA, I. toma in Southeast Asia. Journal of Herpetology 2002. Ecological and phylogenetic correlates of 32: 198–205. feeding habits in Neotropical pitvipers of the FEDER, M. E. AND ARNOLD, S. 1982. Anaerobic genus Bothrops. p. 307–328. In: G. W. Schuett, metabolism and behavior during predatory M. Höggren, M. E. Douglas, and H. W. Greene encounters between snakes (Thamnophis elegans) (eds.), Biology of the Vipers. Eagle Mountain and salamanders (Plethodon jordani). Oecologia Publishing LC., Eagle Mountain. 53: 93–97. MASUNAGA, G., NAKAMURA, Y., AND TOMINAGA, FITCH, H. S. 1960. Autecology of the copperhead. A. 2008. Predation of the Ryukyu brown frog, University of Kansas Publications, Museum of Rana sp., by a neonate of the hime-habu, Natural History 13: 85–288. Ovophis okinavensis, in the northern part of GLAUDAS, X., JEZKOVA, T., AND RODRíGUEZ- Okinawa Island, Japan. Akamata (19): 5–9. (in ROBLES, J. A. 2008. Feeding ecology of the Japanese) Great Basin rattlesnake (Crotalus lutosus, MATSUI, M. 2007. Unmasking Rana okinavana Viperidae). Canadian Journal of Zoology 86: Boettger, 1895 from the Ryukyus, Japan (Amphibia: 723–734. Anura: Ranidae). Zoological Science 24: 199– GREENE, H. W. 1992. The ecological and behav- 204. MORI & TODA—FEEDING CHARACTERISTICS OF PITVIPER 51

MONTEIRO, C., MONTGOMERY, C. E., SPINA, R., ORLOV, N., ANANJEVA, N., AND KHALIKOV, R. SAWAYA, F. J., AND MARTINS, M. 2006. Feeding, 2002. Natural history of pitvipers in eastern and reproduction, and morphology of Bothrops southeastern Asia. p. 345–359. In: G. W. mattogrossensis (Serpentes, Viperidae, Crotalinae) Schuett, M. Höggren, M. E. Douglas, and H. W. in the Brazilian Pantanal. Journal of Herpetology Greene (eds.), Biology of the Vipers. Eagle 40: 408–413. Mountain Publishing LC., Eagle Mountain. MORI, A. AND MORIGUCHI, H. 1988. Food habits POUGH, F. H. AND GROVES, J. D. 1983. Specializa- of the snakes in Japan: A critical review. Snake tions of the body form and food habits of snakes. 20: 98–113. American Zoologist 23: 443–454. MORI, A. AND VINCENT, S. E. 2008. An integrative POUGH, F. H., ANDREWS, R. M., CADLE, J. E., approach to specialization: Relationships among CRUMP, M. L., SAVITZKY, A. H., AND WELLS, K. feeding morphology, mechanics, behaviour, per- D. 2004. Herpetology, Third Edition. Pearson formance and diet in two syntopic snakes. Jour- Prentice Hall, Upper Saddle River. nal of Zoology 275: 47–56. REINERT, H. K. AND CUNDALL, D. 1982. An MORI, A., TODA, M., AND MURAYAMA, N. 2009. improved surgical implantation method for Yearly fluctuation of breeding dates of the radio-tracking snakes. Copeia 1982: 702–705. Ryukyu brown frog, Rana sp., and the Okinawa RODRíGUEZ-ROBLES, J. A. 2002. Feeding ecology tip-nosed frog, Rana narina, over 12 years. of North American gopher snakes (Pituophis Akamata (20): 19–23. (in Japanese) catenifer, Colubridae). Biological Journal of the MORI, A., TODA, M., AND OTA, H. 2002. Winter Linnean Society 77: 165–183. activity of the Hime-Habu (Ovophis okinavensis) SAINT GIRONS, H. 1952. Écologie et éthologie in the humid subtropics: Foraging on breeding des vipéres de France. Annales des Sciences anurans at low temperatures. p. 329–344. In: G. Naturelles, Zoologie et Biologie Animale 14: W. Schuett, M. Höggren, M. E. Douglas, and H. 263–343. W. Greene (eds.), Biology of the Vipers. Eagle SAZIMA, I. 1992. Natural history of the Jararaca Mountain Publishing LC., Eagle Mountain. pitviper, Bothrops jararaca, in Southeastern MUSHINSKY, H. R. 1987. Foraging ecology. p. 302– Brazil. p. 199–216. In: J. A. Campbell and E. D. 334. In: R. A. Seigel, J. T. Collins, and S. S. Brodie, Jr. (eds.), Biology of the Pitvipers. Selva, Novak (eds.), Snakes. Ecology and Evolutionary Tyler. Biology. MacMillan Publishing Company, New SHINE, R. 1991a. Why do larger snakes eat larger York. prey items? Functional Ecology 5: 493–502. NAKACHI, A. 1991. An observation on the stomach SHINE, R. 1991b. Intersexual dietary divergence contents of Trimeresurus okinavensis. Akamata and the evolution of sexual dimorphism in (7): 7–8. (in Japanese) snakes. American Naturalist 138: 103–122. NISHIMURA, M., AKAMINE, H., OYADOMARI, Y., SHINE, R. AND SUN, L.-X. 2003. Attack strategy TAMAKI, H., AND KAMURA, T. 1995. Tracking of of an ambush predator: Which attributes of Habu by radio telemetry 3. Annual Report of the prey trigger a pit-viper’s strike? Functional Habu Study Section, Okinawa Prefectural Insti- Ecology 17: 340–348. tute of Health and Environment (18): 111–124. SHINE, R. AND WALL, M. 2007. Why is intraspecific (in Japanese) niche partitioning more common in snakes than NISHIMURA, M., ARAKI, Y., UEDA, H., AND in lizards? p. 173–208. In: S. M. Reilly, L. B. KAWASHIMA, Y. 1991. Frequencies of prey items McBrayer, and D. B. Miles, (eds.), Lizard Ecology. of Habu, Trimeresurus flavoviridis (Viperidae), Cambridge University Press, Cambridge. in the Okinawa Islands 1. Snake 23: 81–83. THOMPSON, J. N. 2005. The Geographic Mosaic of NOGUEIRA, C., SAWAYA, R. J., AND MARTINS, M. Coevolution. University of Chicago Press, Chicago. 2003. Ecology of the pitviper, Bothrops moojeni, TODA, M., YAMAMOTO, R., AND KADOTA, Y. 2003. in the Brazilian Cerrado. Journal of Herpetology Terrestrial herpetofauna of Gushikawajima and 37: 653–659. Yanahajima Islands of Izena Village, the Okinawa 52 Current Herpetol. 30(1) 2011

Island Group. Biological Magazine of Okinawa (Viperidae: Ovophis okinavensis): Key roles for 41: 33–41. (in Japanese with English abstract) head size and body temperature. Biological VALDUJO, F. H., NOGUEIRA, C., AND MARTINS, M. Journal of the Linnean Society 93: 53–62. 2002. Ecology of Bothrops neuwiedi pauloensis VINCENT, S. E., DANG, P. D., HERREL, A., AND (Serpentes: Viperidae: Crotalinae) in the Brazilian KLEY, N. J. 2006. Morphological integration Cerrado. Journal of Herpetology 36: 169–176. and adaptation in the snake feeding system. A VINCENT, S. E. AND HERREL, A. 2007. Functional comparative phylogenetic study. Journal of and ecological correlates of ecologically-based Evolutionary Biology 19: 1545–1554. dimorphisms in squamate reptiles. Integrative and Comparative Biology 47: 172–188. VINCENT, S. E. AND MORI, A. 2008. Determinants Accepted 10 June 2011 of feeding performance in free-ranging pit-vipers